On Thu, Mar 16, 2017 at 05:08:51PM +0800, Qu Wenruo wrote:
> Btrfs balance will reloate date extent, but its hash is removed too late
^^^^^^^ relocate
> at run_delayed_ref() time, which will cause extent ref increased
> during balance, cause either find_data_references() gives WARN_ON()
> or even run_delayed_refs() fails and cause transaction abort.
>
> Add such concurrency test for inband dedupe and data balance.
>
> Signed-off-by: Qu Wenruo <quwenruo@xxxxxxxxxxxxxx>
> ---
> tests/btrfs/202 | 109 ++++++++++++++++++++++++++++++++++++++++++++++++++++
> tests/btrfs/202.out | 3 ++
> tests/btrfs/group | 1 +
> 3 files changed, 113 insertions(+)
> create mode 100755 tests/btrfs/202
> create mode 100644 tests/btrfs/202.out
>
> diff --git a/tests/btrfs/202 b/tests/btrfs/202
> new file mode 100755
> index 0000000..60bb924
> --- /dev/null
> +++ b/tests/btrfs/202
> @@ -0,0 +1,109 @@
> +#! /bin/bash
> +# FS QA Test 203
> +#
> +# Btrfs inband dedupe with balance concurrency test
> +#
> +# This can spot inband dedupe error which will increase delayed ref on
> +# an data extent inside RO block group
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2017 Fujitsu. All Rights Reserved.
> +#
> +# This program is free software; you can redistribute it and/or
> +# modify it under the terms of the GNU General Public License as
> +# published by the Free Software Foundation.
> +#
> +# This program is distributed in the hope that it would be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program; if not, write the Free Software Foundation,
> +# Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
> +#-----------------------------------------------------------------------
> +#
> +
> +seq=`basename $0`
> +seqres=$RESULT_DIR/$seq
> +echo "QA output created by $seq"
> +
> +here=`pwd`
> +tmp=/tmp/$$
> +status=1 # failure is the default!
> +trap "_cleanup; exit \$status" 0 1 2 3 15
> +
> +_cleanup()
> +{
> + cd /
> + kill $populate_pid &> /dev/null
> + kill $balance_pid &> /dev/null
> + wait
> + # Check later comment for reason
> + $BTRFS_UTIL_PROG balance cancel $SCRATCH_MNT &> /dev/null
> + rm -f $tmp.*
I sometimes saw "btrfs balance start" hold $SCRATCH_MNT from umounting
and result in fs inconsistency after test. Some debug codes show that
the "btrfs balance start $SCRATCH_MNT" is still running. If I add what
btrfs/061 does:
while ps aux | grep "balance start" | grep -qv grep; do
sleep 1
done
test finishes fine. But btrfs-balance(8) says balance cancel should
block and wait for the backgroud balance process to finish. Seems like a
"balance cancel" bug?
Otherwise these three patches look fine to me overall, except the wanted
"dedupe feature check", and btrfs/200 is quick enough to fit in 'quick'
group.
I think the "dedupe feature check" should _notrun the tests if current
btrfs has incompatible features set, e.g. compress and/or nodatacow.
Thanks,
Eryu
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
